The Enrico Fermi reactor was a sodium-cooled fast breeder demonstration re­actor, producing 200 MW(t) [61 MW(e)]. The plant was located near Lagoona Beach, Michigan, and started operation in 1963. After extended low-power op­eration, power raising took place during 1966. When this was being done, it was noted that the coolant temperatures above 2 of the 155 fuel assemblies (clusters of fuel rods) were higher than normal and the temperatures above an­other assembly were lower than normal.

The reactor was shut down, and the fuel assemblies were rearranged in the core to determine whether these abnormal temperatures were dependent on location in the core or were characteristic of the fuel assemblies themselves.

On October 5, 1966, the rise to the selected power level [67 MW(t)] for these tests on the rearranged fuel elements was begun. At about 3 p. m., with the re­actor at a power level of 20 MW(t), the reactor operator observed a control sig­nal, indicating that the rate of change of neutron population was erratic. The problem had been experienced before and was thought to be due to random electrical fluctuations in the control system. The reactor was placed on manual control, and when the instability disappeared, automatic control was again se­lected and the increase in power resumed.

At 3:05 p. m., with the reactor power at 27 MW(t), the erratic signal was again oh — selved. Shortly after that it was noted that the control rods were withdrawn farther than normal. A check of the core exit temperatures showed that the outlet tem­peratures from two subassemblies were abnormally high at 380 and 370°C (715 and 695°F), compared with a mean bulk outlet temperature of 315°C (600°F).

At 3:09 p. m., alarms occurred from the ventilation monitors in the upper building ventilation exhaust ducts. The building was automatically isolated—no one was inside at the time—and a radiation emergency was announced. The re­actor power increase was stopped at 31 ^W(t), and a power reduction was started. By 3:20 p. m., the power had decreased to 26 ^W(t) and the reactor was manually tripped and shut down indefinitely. Over the next year, many of the assemblies were removed and examined, and it was found that the bulk of the fuel in two of the fuel assemblies had melted. It was not until the end of the ex­amination period that the cause of the accident was discovered. The cause was relatively trivial. Below the core, six small Zircaloy plates had been installed to guide the flow of sodium into the upward direction. One of these Zircaloy plates had broken loose and blanked off the entry to a few subassemblies, causing almost total flow starvation.

The damage to the reactor was repaired with a specially designed remote handling tool, and the reactor reached full power output again in October 1970, 4 years after the accident.

Although the Enrico Fermi accident led to no injury or release of activity out­side the containment shell, 10,000 curies of fission products were released to the circulating sodium coolant. The accident focused attention on the potential problems of flow blockages caused by foreign bodies within the circulating sodium. In particular, any part of the reactor that may be susceptible to vibration damage, causing the release of foreign material, must be carefully evaluated. In the design of modern reactors, very thorough flow testing of the various com­ponents is carried out. It is noteworthy that the zirconium plates were added at a very late stage in the design and may not have had the same level of quality assurance as the other components in the Enrico Fermi reactor. Late “fix-ups" of this kind and of the kind that occurred at Hunterston must be avoided.

The damage to the fuel assemblies did not propagate to adjacent fuel assem­blies, and the evidence from this incident that the accident did not escalate was encouraging.